Dementia has reached epidemic proportions, with an estimated 4.6 million new cases worldwide each year. With an aging world population, the prevalence of dementia will increase dramatically in the next few decades. Of the predicted 114 million who will have dementia in 2050, about three-quarters will live in less developed regions. Although strongly age-related, dementia is not an inevitable part of aging but is a true disease, caused by exposure to several genetic and nongenetic risk factors. Prevention will be possible when the nongenetic risk factors have been identified. Apart from age, more than 20 nongenetic risk factors have been postulated, but very few have been established by randomized intervention studies. Elevated blood concentrations of total homocysteine and low-normal concentrations of B vitamins (folate, vitamin B12, and vitamin B6) are candidate risk factors for both Alzheimer's disease and vascular dementia. Seventy-seven cross-sectional studies on more than 34,000 subjects and 33 prospective studies on more than 12,000 subjects have shown associations between cognitive deficit or dementia and homocysteine and/or B vitamins. Biologically plausible mechanisms have been proposed to account for these associations, including atrophy of the cerebral cortex, but a definite causal pathway has yet to be shown. Raised plasma total homocysteine is a strong prognostic marker of future cognitive decline, and is common in world populations. Low-normal concentrations of the B vitamins, the main determinant of homocysteine concentrations, are also common and occur in particularly vulnerable sections of the population, such as infants and elderly. Large-scale randomized trials of homocysteine-lowering vitamins are needed to see if a proportion of dementia in the world can be prevented.